Abstract: This paper describes the process and characteristics of the great floods occurred in the Changjiang River, the Nenjiang River and the Songhua River during 1998. The flood of the Changjiang River is compared with that of 1954, and the reason of higher water level under same discharge is analyzed in detail. Some considerations on countermeasures are discussed. A lot of data calculated in this paper are very useful for scientific research and practice on flood control.
Key words: Changjiang(Yangtze) River, Nenjiang River, Songhua River, 1998 Flood
I. INTRODUCTION
In 1998, the heaviest flood since 1954 occurred in the whole Changjiang(Yangtze) River Basin, the heaviest flood in history in the Nenjiang River Basin and the flood with the frequency of over 100 years in the Songhua River Basin. The deep going analyses of these big floods are important for flood prevention, project planning and disaster mitigation.
II. FLOOD IN CHANGJIANG RIVER BASIN
1. Precipitation in 1998
Fig.1 shows the distribution of monthly precipitation in the Changjiang River Basin. In June, there was much rain to the south of the Changjiang River. It caused the early flood in Boyang and Dongting Lakes as described in the following section. Rain belt was located along the middle and lower reaches of Changjiang River in July, and the upper reach of Changjiang River received much rain in August.
Fig.2 shows the daily precipitation and cumulative one at selected stations. Total amount of precipitation during JJA(June, July and August) at Yichang is 583mm. At around DOY(Day Of the Year)180(29th, June), there was an event with 127mm of precipitation during seven days, and 254mm was precipitated during DOY 198(17th, July) and 233(21th, August). A large event was held during DOY 237(25th, August) and 242(30th, August) with total amount of 129mm.
In Wuhan, total precipitation was 835mm during JJA, especially precipitation in July reached to 729mm. An event during DOY 202(21st, July) and 206(25th, July) is the largest with 471mm of total precipitation, and followed by 97mm event during DOY 210(29th, July) and 214(2nd, August).
The events during DOY 163(12nd, June) and 180(29th, July) are remarkable at Changsha located upstream of Dongting Lake. At Nangchang upstream of Boyang Lake, two events at around DOY 170(19th, June) and 210(29th, July) are clear. The total amounts of precipitation of these events are 550mm and 410mm, respectively. These events contributed to the expansion of water area of Dongting and Boyang Lakes and high water level in the middle to lower reach of Changjiang River.
2. General Situation
From 28th June to 3rd September, eight successive flood events occurred in the Changjiang River. The hydrographs at Yichang, Shashi and Hankou hydrological stations are shown in Fig. 3.
From 11th to 28th June, it was Meiyu (Bai-u) rain period. Intensive rainfall belts located from east to west and centered in Dongting Lake, Boyang Lake, especially in Xinjiang River Basin, a branch of Boyang Lake. The average cumulative rainfall in northern part of Jiangxi Province was 500 mm, and over 700 mm at 10 rain gage stations, with the extraordinary value of 1,115 mm at Shangqing Rain Gauge Station. The maximum daily rainfall was 271 mm in Jindezhen City. From 28th on, water levels of all hydrological stations downstream of Jianli Station exceeded warning levels respectively.
The first flood event occurred during the period from 28th June to 10th July. Discharge at Yichang Station reached 53,500 m3/s. Water levels at Jianli, Wuxu and Quijiang exceeded the highest values in history.
From 10th July on, rainfall moved to upper basin and the Hanjiang River Basin, second flood event lasted until 20th July. The discharge was 56,400 m3/s at Cuntan Station and over 50,000 m3/s at Yichang Station.
The third flood event was from 21st to 31st. Heavy rainfall fell along middle and lower river reach for a long time. Floods from upper reach and Dongting Lake met in the middle reach and then joined flood from Boyang Lake on the way of propagation downward. So, after 25th July, the water levels at Shishou, Jianli, Wuxi, Jiujiang, Chenglingji and Hukou created new historical records.
From 1st to 10th August, the flood from upper reach met that from local area of the Three Gorge, the fourth flood event was formed. The discharge of Yichang Station was 61,500 m3/s on 7th and the water level of Shashi was 44.95 m on 8th , being 0.28 m higher than that in 1954.
Fifth flood event was from 10th to 15th August. It was caused by the flood from upper basin, the Qingjiang River Basin and local area of the Three Gorge. On 13th August, the discharge and water level at Yichang Station reached 62,800 m3/s and 54.03 m respectively.
The sixth flood event was the most serious one. From 16th to 23rd August, the flood from upper basin, Qingjiang River Basin, Dongting Lake and Hanjiang River Basin encountered in the middle river reach of Changjiang River. The discharge at Yichang Station was 63,600 m3/s on 17th. The water level was 45.22 m at Shashi Station on 18th, being 0.55 m over the historical record.
The seventh flood event from 24th to 29th August was formed by the water from upper basin, the Gezhoubar, Geheyan and Zhanghe Reservoir played an important role in reducing the flood peak. The peak discharge of Yichang Station was 56,300 m3/s on 26th August.
The eighth flood event from 30th August to 3rd September was also from upper basin. On the 1st September, the inflow of Gezhoubar Reservoir was 58,800 m3/s. After reduction of the flood peak, the discharge at Yichang Station was 57,400 m3/s.
From 4th to 26th September was the recession period. The discharge and water level reduced gradually. On 26th September, the water levels of all hydrological stations were below warning levels.
3. Characteristics of the Flood in the Changjiang River Basin
1) Early rise of water level
In 1998, the base flow in river course and reservoirs located on the middle and lower basin of the Changjiang River was very high, two flood processes even occurred in March. The water levels of Xiangjiang, Ganjiang and Wuhe rivers, branches of Dongting Lake and Boyang Lake, once exceeded respectively the warning levels in March. For all the hydrological stations along the middle and lower river reaches, the starting dates of water level rising and exceeding warning level were earlier than other years.
2) Rainfall with high intensity, long duration and wide covering area
The rainfall center was located in the Dongting Lake and Boyang Lakes from 11th to 28th June, in the Three Gorge area from 28th June to 1st July, in upstream tributaries as the Jingsha River, the Mingjiang River and the Jialing River in second ten-day of July, in nearly whole basin, especially Wuhan City, Lishui River and Yuanjiang River of Dongting Lake, Wujiang River of upper basin in the last ten-day of July, in upper stream branches as Mingjiang River and Wujiang River at the beginning of August, and in the Three Gorge area after the third of August. It can be seen from the above that the rainfall center covered wide areas during different periods.
3) Rapid rise of water level
Water level in main river course rose rapidly. For example, at the Hankou Hydrological Station, water level rose 0.8 m on 27th June, and 3 m from 25th to 30th June.
4) High water level with long duration
Water levels at Juijiang and Hukou stations exceeded the warning level from 24th June when to 26th September when water levels of all stations were below warning levels, with high water level being kept for more than three months.
Table 1 shows the days when water levels exceeded corresponding warning levels and the highest levels in history for different hydrological stations along the Changjiang River.
5) Superposition of floods
In 1998, the floods from upstream, Boyang Lake, Dongting Lake, Three Gorge area and Hanjang River Basin encountered several times.
The floods from Dongting Lake and Boyang Lake are superposed during second and third ten-day of June, those from upper basin and middle basin during first and second ten-day of July, from upper basin, Boyang Lake and Dongting Lake during third ten-day of July, from Hanjiang River Basin and main river course during the first and second ten-day of July, from upper basin, Three Gorge area and Qingjiang River Basin during the first ten-day of August, from Hanjiang River and its branch during the second ten-day of August, from upper basin, Three Gorge area and Qingjiang River Basin on 16th August.
4. Comparison with the Flood in 1954
The flood occurred in 1954 is the biggest one since the foundation of People's Republic of China. It is sutable to compare the 1998 flood with the 1954 flood to understand the characteristics of the 1998 flood.
Tables 2 to 6 show the comparison between 1954 and 1998 floods. These tables explain following things:
- The total average rainfall amount over the whole basin and the rainfall amounts of most individual stations in 1998 are less than those in 1954(Table 2).
- The comparison of maximum 30 days flood volumes in river course for Yichang and Hankou stations are nearly the same(Table 3).
- The peak discharges at the majority of the stations in 1998 are less than those in 1954(Table 4).
- The highest water levels of most stations in 1998 are higher than those in 1954(Table 5).
- Under the same discharge as 1954, the water levels in 1998 are higher than those in 1954(Table 6).
The duration of high water level in 1998 is longer than that in 1954. It can be concluded from above data that the water levels are higher than those in 1954 under the same discharge.
5. Satellite Monitoring of the Flood
The inundated condition was taken by NOAA/AVHRR satellite. Figs.4 and 5 show the near-infrared image of Dongting and Boyang Lakes. Threshold value which divides inundated and non-inundated area is selected by manual inspection for each image because unique value could not be determined due to various conditions such as turbidity of water, atomospheric state and so on. Flooded area can be seen as dark because water absorbs infrared light.
Although image of Dongting Lake in August 1997 after rainy season shows relatively large water area, water area was decreased in September 1997 and May 1998 images. In July image, expansion of water area in Dongting Lake was recognized. After August image, the expansion of main channel of Changjiang River is recognized.
In Fig.5, enlargement of the water area in Boyang Lake was recognized after July1 image. This was caused by the early rain and it keeps high water level in the lower reach of Changjiang River.
6. Causal Analysis
Higher water levels under same discharges at some stations are certainly caused by raised river bed due to sedimentation. But it is important to know that it is caused by the increase of incoming sediment to the main river course or by the movement of sediment in the river course itself. The Cuntan Hydrological Station near Zhongqing City can monitor sediment from the Jialing River. Yichang Hydrological Station is the starting point of the middle reach of Changjiang River, so sediment from upper basin can be monitored there. It can be seen from Table 7 that sediment from upper basin ( Shichuan Province and Yunnan Province ) does not increase so much although forest covering area is reduced from 30% to 13% in Shichuan Province and from 55% to 22% in Yunnan Province since 1930 due to destruction of forest. Of course, the destruction of forest may result in more sediment, earlier and higher flood peak. But the increased sediment is actually interrupted by small and middle-sized reservoirs constructed on branches.
Now that the sediment entering main river course is not increased, why the water levels are higher than before under the same discharge. After further analysis, it can be found that main reasons are decrease of lake area connecting the Changjiang River, variation of relation between Dongting Lake and Changjiang River and variation of scour and filling in main river course.
1) Reduction of lake area
Historically, there have been many lakes connecting the Changjiang River and diverting water and sediment during flood season. But the total area of these lakes is reduced from 17,200 km2 in 1825 to less than 6,600 km2 in 1990 due to sedimentation. For example, the number of lakes in Hubai Province is reduced from 1,066 to 309. As for lake decline, the Dongting Lake is the most typical example. Table 8 shows the variation of its area and storage capacity.
Decline of the Dongting Lake is caused by sediment from its branches and enclosure of the lake to be farmland. Since 1949, sediment increases 0.1 billion m3 every year and the elevation of lake bottom is raised 1.7 m. The area of lake enclosure to be farmland was 13.2 km2 per year from 1825 to 1949, 87 km2 per year from 1949 to 1954 and 193.5 km2 per year from 1954 to 1998. The decline of lake makes the occurrence of flood in the area surrounding lake more frequent, being once every 5 years from 1950 to 1957, once every 3 or 4 years from 1970 to 1990 and twice every 3 years after 1990.
The decline of lakes reduces their capacity of diverting water and sediment, more sediment has to be drained from the Changjiang River.
2) Variation of the relation between the Changjiang River and the Dongting Lake
As mentioned before, the Dongting Lake used to play the role of diverting water and sediment from Changjiang river during flood season through four rivers connecting the river and the lake, namely, Taibingkou, Songtzekou, Tiaoxianhou and Outzekou. In 1931, the diversion ratio was 50.4%, but reduced to 20% at present, corresponding to the increase of 763 m3/s on discharge and 15% on sediment in Changjiang River.
The reduction of diversion ratio is caused by the water head decrease of these four rivers. The river bed and water levels at outlets of these rivers to Dongting Lake become higher due to sedimentation, while those at intakes from Changjiang River become lower due to following two reasons:
For flood control, the meandering river reach downstream of Jingjiang was cut to be straight since 1967, the river course has been shorten 76 km and flow velocity increases accordingly. It is favorable to flood wave propagation, but also scours the river bed downstream of Jinjiang. Besides, the operation of Gezhoubar Reservoir may also scour the river bed downstream of Yichang. The intakes of four rivers mentioned above are located along the scouring river reaches, so the river bed and water levels at these intakes become lower.
The river bed near the Jianli Hydrological Station located in this river reach is scoured, but its water level is still higher than that in 1954 because the discharge increases due to the reduction of diverting flood volume to Dongting Lake.
3) Sedimentation in river reach between Chenglingji and Wuhan
The scoured sediment is moved to and filled in the river reach between Chenglingji, the outlet of the Dongting Lake to the Changjiang River, and Wuhan City. Through the comparison on measured sediment concentration between input and output, it can be known that from 1959 to 1986, 0.5784 billion m3 of sediment filled there, the average thickness of increased sedimentation is 1.41 m, corresponding to 0.05 m every year. For example, the cross section at the Luosan Hydrological Station has been reduced nearly 6,000 m2, resulting in the rise of water level from 1 to 1.5 m depending on different discharges. Fig.6 is the variation of its cross section from 1954 to 1986.
After the completion of the Three Gorge Reservoir, the river bed downstream of the dam will be certainly scoured significantly. Where the scoured sediment will be moved to and filled is a very important problem. The fact that higher water level under the same discharge in 1998 reveals this problem which should be carefully taken into account in advance.
III. FLOOD IN THE NENJIANG AND THE SONGHUA RIVER BASIN
1. Precipitation in 1998
Fig.7 shows the monthly precipitation in the Northeast China. It is apparent that Keshan and Bugt received much rain in July. The area surrounded by Harbin, Bugt and Keshan had much precipitation in August.
Fig.8 shows the daily precipitation in Keshan, Bugt and Harbin. The two events at around DOY 210(29th, July) and 220(8th, August) are clear. Total precipitation during this period is 457mm. Bugt had continuous rain after DOY 160(9th, June), and 142mm event at around DOY 200(19th, July) and 154mm event at around DOY 220(8th, August) are remarkable.
2. General Situation
The Nenjiang River Basin is one of three storm centers in the Northeast China, with the annual runoff volume of 25 billion m3. In 1998, the rainy season of the Nenjiang River Basin came earlier than in normal years, the first storm with 68 mm occurred on 28th May. In normal years, the annual precipitation of this basin is 300 to 400 mm. But, the average rainfall from June to August in 1998 over the whole basin is 486 mm, being 63% more than normal years. The total runoff volume in July and August is 38 billion m3. Four flood events occurred successively. The second and the third flood events were the heaviest ones in history. The Nenjiang River is the upstream tributary of the Songhua River. The heaviest flood in Nenjiang River Basin resulted in the heavy flood with the frequency of occurrence of over 100 years in the Songhua River. The discharge and water level exceeded the historical record. Fig. 9 indicates the hydrograph at the Harbin Hydrological Station and Table 9 shows the situation of some hydrological stations along the Nenjiang River and Songhua River.
The embankment was breached in three places along the Nenjiang River, left dike in Larhai on 13th August, right dike in Tailai on 15th August and left dike in Darduopao on 17th August. The embankment of Songhua River was also breached on one place located in Shaoyuan County. Part of the oil wells of the Daqing Oil Field and the railway to Qiqiharl were inundated. These breaches resulted in considerable economic loss, but also mitigated the pressure on downstream large cities as Harbin and Jiamose.
3. Satellite Monitoring of the Flood
Fig.10 shows the NOAA/AVHRR infra-red images on 23rd/Aug./1997 and 20th/Aug./1998.
In 1997 image, river courses of the Nenjiang and the Songhua river are clearly seen. Along the Nenjiang, the width of water area(low infra-red reflectance) looks wide. It includes the wetland along the Nenjiang River. Songhua River can be recognized as thin line on the image.
On 20th/Aug./1998, the water level was nearly the crest of flood hydrograph at Harbing. River channels are expanding as much as several tens of kilometers, and waterlogged area were seen anywhere in the basin.
4. Causal Analysis
The serious flood disaster occurred in the Nenjiang River Basin is mainly due to lack of structural measures. The Songhua River has two major branches, the Nenjiang River and the Second Songhua River. On the latter, there are Baishan and Fongman reservoirs which can control the flood with frequency of one to 100 years, but there is no large control project on the Nenjiang River, also the standard of embankment of this river is very low. On both sides of the Nenjiang River, there are a lot of marshland and depressions with the total area of 1,400 km2. In these areas, the population is very small and economy is quite substandard. If flood diversion basins are constructed in these areas, the economic loss in case of the flood during 1998 would have been reduced greatly.
Besides, the forest covering area of Daxinganling and Xiaoxinganling is reduced from 70-80% in 1949 to 30% at present. It results in rapid flow concentration, soil erosion and sedimentation in river course.
IV. SOME CONSIDERATION ON FLOOD PREVENTION
The heavy flood in 1998 was of course caused by abnormal climate, especially the influence of El Nino and La Nina. But the negative effect of some human-being activities should be introspected, such as enclosure of lakes to make farmland, destruction of forest and so on. Both structural and non-structural countermeasures should be taken into account. Apart from water and soil conservancy, afforestation, upgradung of flood-control standard, reinforcement of embankment, construction of more reservoirs in upper basins, transformation of slope fields to be terraced fields, returning farmland to lake, dredging of river course, improvement of weather and flood forecasting, application of remote sensing and geographic information system, completion of flood prevention law, promoting the flood prevention insurance. This paper will mainly discuss the following countermeasures.
1) Increase of investment to water resources sector
With rapid development of society and economy, the economic loss resulting from flood and waterlogging disaster increases significantly. It was 50 billion yuan ( RMB ) in 1991, 170 billion yuan in 1994, 150 billion yuan in 1995, 220 billion yuan in 1996, while 270 billion yuan in 1998.
Although water resources is specified as the first basic estate of China, but the investment to this sector is less than some other ones. From 1991 to 1997, the total investment to water resources sector is 115 billion yuan, being 1% of the total investment of the country and with annually increasing rate of 21%; while those are 1391.058 billion yuan, 12.4% and 32.29% for energy sector; 879.9 billion yuan, 7.85% and 47.4% for transportation sector; 475.2 billion yuan, 4.2% and 65.5% for communication sector; 439.818 billion yuan, 3.9% and 33.5% for city construction. It can be clearly seen from above figures that the investment to water resources estate should be greatly increased.
2) Further knowledge of regularities of sediment transportation in lake and river course, especially the Changjiang River
As mentioned above, the high water level in Changjiang River is mainly caused by the variation of scour and filling in river course and it would be a noticeable problem after the construction of the Three Gorge Dam. So it is very important to study the regularities of sediment transportation. There was not heavy flood in Changjiang River over 40 years. It hides this problem and also the lack of knowledge on sediment transportation. It is the time to recognize and study this problem again.
3) Adjustment of operational scheme of the Three Gorge Reservoir
According to the design scheme, the reservoir will begin to store flood water when the inflow of reservoir exceeds 56,700 m3/s. That means in case of the floods during 1998, the reservoir would be seldom used. It is evidently not reasonable. For eight flood events in 1998, the Three Gorge Reservoir at least can reduce the peak discharge to 45,000 m3/s for seven flood events except the last one. The Three Gorge Reservoir can play very important role to flood prevention of the Changjiang River. So the operational scheme should be studied again combining with the sediment problem in downstream river course and reservoir area.
4) Peaceful coexistence with flood-construction mode of flood diversion basin
The occurrence of flood and waterlogging will be inevitable for a very long time, especially in China. It is important to learn how to peacefully coexist with flood. China has a large population and shortage in farmland. It is impossible to leave corridor along river as the way for big flood passage. Besides, it is impossible to construct large reservoir in plain area either. Therefore, a lot of flood diversion basins were established. But at present, it is difficult to use them due to rapid development of economy in these basins. In order to peacefully coexist with flood, that means these basins can be used with the economic loss reduced to the minimum, the construction mode of flood diversion basin should be considered. What follow are some examples. A house can be designed with long legs, there are only some solid pillars in the first floor, to make the room for flood water. A basin can be divided into several sub-basins in order to avoid that the basin would be fully inundated once it is used although the diverting volume is not so large. For agriculture in these basins, it is better to select the crops with high stem and harvest time before flood season. Industry is not suitable to be developed in these areas. High ground platform should be constructed in the basin for temporal refuge.
5) Construction of the Nierji Reservoir and flood diversion basins in the Nenjiang River Basin
As mentioned before, one of main reasons of flood disaster in the Nenjiang River Basin is lack of large flood-control project. Nierji, 130 km upstream of Qiqiharl, is an ideal place for constructing a large-sized reservoir. It can control 20% of total watershed area. The planned storage capacity is 1.6 billion m3. In case of 1998, it can reduce 5,000 m3/s of peak discharge in Qiqiharl, 4,500 m3/s in Daqing and 0.4 m of water level in Harbin. It can protect 89 towns, 676 villages, 586,000 houses, 317,666 ha farmland, 300,866 ha grassland and more than 700,000 of population. Besides, it can improve irrigation and navigation conditions greatly. Such a control project should be constructed as early as possible.
But in case of 1998, after the construction of Nierji Reservoir, the downstream peak discharge would be 11,400 m3/s, still exceeding the insurance discharge of 7,790 m3/s. So the establishment of several flood diversion basins with total storage capacity of 3 billion m3 is also necessary. The principle of flood control for the Nenjiang River is to intercept flood in upper basin, divert flood in middle basin and drainage flood in downstream river course.
V. FINAL REMARKS
Floods of China in 1998 caused significant loss in economy but it is also a good lesson. It reveals some problems which had not been recognized enough. Deep going analysis on it is very important for flood-prevention planning and practice in China, as well as for scientific research in many fields.
ACKNOWLEDGEMENT
One of the authors, Prof. Li Jiren, stayed CEReS, Chiba University as a foreign researcher after the flood. This paper is one of outcomes during the stay. We appreciate Prof. Y. Yasuda, director of CEReS, and colleagues to give a chance to complete the paper.
Other author, Dr. A. Kondoh got a chance to visit the Changjiang River to understand actual situation of the 1998 great flood. We express our gratitude to Prof. K. Takeuchi, Yamanashi University, who gave a valuable opportunity to visit China.
Finally, we defer to the people who make an harsh effort to defend the bank of Changjiang River, and express our sincere sympathy to the victims of the flood.
REFERENCES
Han Qiwei(1998): Analysis of reason on water level exceeding historical record in middle reach of Changjiang River. Journal of China Institute of Water Resources & Hydropower Research, (Printing).
Global Surface Summary of Day Version 5 (1999): http://www.ncdc.noaa.gov/cgi-bin/res40.pl?page=gsod.html. National Climatic Data Center/National Oceanic and Atmospheric Administration.